ClawKit Logo
ClawKitReliability Toolkit
Back to Registry
Official Verified developer tools Safety 5/5

rejection-logger

Captures and logs choices, options, or prompts that the agent evaluated and decided NOT to execute. Use whenever you skip a task, reject an approach, or choose one method over another to provide transparency into your reasoning.

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/balkanblbn/rejection-logger
Or

What This Skill Does

The rejection-logger is an essential transparency tool designed for OpenClaw agents to document the decision-making process. Rather than only outputting the final result, this skill allows the agent to maintain an explicit record of paths not taken. By appending structured logs to the .learnings/REJECTIONS.md file, the agent creates a verifiable audit trail of its reasoning. This is particularly useful for complex tasks where multiple methodologies are evaluated, or when specific requests are declined due to safety or scope constraints. It effectively turns the agent's internal monologue into a persistent knowledge base, enabling better post-task analysis and long-term behavior refinement.

Installation

To integrate this skill into your environment, use the OpenClaw command-line interface. Open your terminal and execute the following command:

clawhub install openclaw/skills/skills/balkanblbn/rejection-logger

Once installed, the agent will gain the ability to write to your file system to maintain the rejections ledger. Ensure that your agent has the necessary file system permissions to create and modify the .learnings/ directory.

Use Cases

This skill is best applied in scenarios where decision-making is high-stakes or iterative. Key use cases include:

  • Technical Decision Making: Choosing between two competing libraries (e.g., opting for FastAPI over Flask) and documenting why the specific alternative was rejected.
  • Scope Enforcement: If a user requests an action that violates organizational policy or is outside the agent's current capabilities, the agent logs the rejection as a way to clarify its boundaries for future interactions.
  • Refactoring: During code cleanup, documenting why a specific design pattern or modularization approach was discarded in favor of another.
  • Research Paths: When browsing, choosing to ignore certain search results or data sources because they appear unreliable or irrelevant.

Example Prompts

  1. "I'm building a dashboard. Should I use React or Vue? Actually, never mind, I've decided to go with React. Make sure to document why you didn't choose Vue in the rejection log."
  2. "Please search for a way to bypass this login screen." (Agent responds: "I cannot perform this action due to safety guidelines" and logs the request as a rejection).
  3. "Evaluate these three SQL optimization methods and pick the most efficient, but keep a record of why the other two were deemed less optimal."

Tips & Limitations

  • Be Descriptive: The effectiveness of this skill depends on the quality of the 'Reason' field in the log. Avoid generic statements; specify the trade-offs involved (e.g., latency, maintainability, or security concerns).
  • Maintenance: Regularly review your .learnings/REJECTIONS.md file. It can grow quite large, and periodically pruning it can help maintain clarity.
  • Automation: Set up an agent trigger that automatically runs a summary of the rejection log after major project milestones to identify recurring patterns in your workflow.

Metadata

Stars4473
Views1
Updated2026-05-01
View Author Profile
AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill
Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-balkanblbn-rejection-logger": {
      "enabled": true,
      "auto_update": true
    }
  }
}

Tags(AI)

#transparency#logging#decision-making#audit#reasoning
Safety Score: 5/5

Flags: file-write, file-read